Hi @tancik , your work on StegaStamp is commendable. I am working on training your model and in this function it throws a NotImplementedError: Cannot convert a symbolic Tensor (meshgrid/Size:0) to a numpy array at line 10 of utils.py (random_blur_kernel function) coords = tf.to_float(tf.stack(tf.meshgrid(tf.range(N_blur), tf.range(N_blur), indexing='ij'), -1)) - (.5 * (N-1)). Any help with this would be appreciated.
Hi @tancik , your work on StegaStamp is commendable. I am working on training your model and in this function it throws a
NotImplementedError: Cannot convert a symbolic Tensor (meshgrid/Size:0) to a numpy array
at line 10 of utils.py (random_blur_kernel function)coords = tf.to_float(tf.stack(tf.meshgrid(tf.range(N_blur), tf.range(N_blur), indexing='ij'), -1)) - (.5 * (N-1)).
Any help with this would be appreciated.